It's preprint day! (a well deserved week-end now!) 🥳
A reduced-order model for parametrized Optimal Transport problems - Inria - Institut national de recherche en sciences et technologies du numérique inria.hal.science/hal-05587825

It’s preprint day! 😏

It’s preprint day! With Q. Merigot and our phd A. Cances! 🙌

Y a pire pour son jogging (et une conf)

We are recruiting four positions connected to Machine Learning, Statistical Learning, and AI for Science in the Applied Mathematics department at École polytechnique. Join our vibrant community at IP Paris and Hi! Paris IA center. List below🧵 tinyurl.com/3jpw9t26

😌

Tres sympa lire l’évaluation du jury IUF (surtout quand on est lauréat, j’avoue) et première fois dans ma vie que le seule point faible est “a bit too trendy for my taste”…si seulement j’étais si trendy que ça 🤣

Onze personnalités liées à l’Université Paris-Saclay nommées à l’Institut universitaire de France (IUF) !

Une reconnaissance prestigieuse pour leur engagement et la qualité de leurs travaux de recherche.

🔗 tinyurl.com/yey44wx2

IUF 🍾👍

Le dossier est trop jeune
Le dossier est trop théorique
Le dossier est trop numérique
Le dossier est trop “physique quantique”
Les dossier est trop “math économique”
Bon on peut se mettre d’accord ou pas? 🤣

C’était une campagne de (pas) recrutement prof assez traumatisante 😅 il me faudra une année pour me reprendre 🤣

Dual cone and polar cone are sets in the dual of a vector space. They generalize the notion of orthogonal subspace.

Voronoi diagram partitions the space in a proximity graph (cells) according to some sites. Each cell is the portion of the space where every points are closer to a given site than any other sites. cs.tau.ac.il/~michas/ak-v...

Se la peter un petit peu 🤏🤣

Ca fait un peu peur…😔

Bon y’a pas beaucoup des postes PR mais au moins y en a deux ou dans le profil y a “transport optimal”…je pense que c’est quand même un record😅

The Hausdorff distance quantifies how far two sets are from each other by measuring the largest minimal distance needed to cover one with the other. It captures worst-case deviation: if one set has points far from the other, the distance is large.

Retailleau grillé par un magistrat en direct sur Inter qui lui rappelle que LR a voté l'article qui a empêché l'exécution de l'OQTF visant l’influenceur algérien "Doualem". "Ca m'étonne", balaye le ministre. Salamé laisse passer.

I didn’t think Trump’s second act would be that bad. Oh, how wrong I was | Arwa Mahdawi The Maga kingpin has only been president for a couple of weeks, but the chaos he has unleashed is already much worse than I thought it could be. I’m genuinely frightened The day after Donald Trump won the election, everyone in my liberal corner of…

I didn’t think Trump’s second act would be that bad. Oh, how wrong I was | Arwa Mahdawi

AHAHAHAHAHAHAHAHAHAHAHAHAHAHAH kind of “gnegnegnegne they cheated gnegnegngne”

I’ve just done my #eXit! the X exodus is massive. Don't lose any of your followers. Thanks to #HelloQuitX I've registered 592 new passengers for a journey to #BlueSky & #Mastodon. Join us on app.helloquitx.com and automatically find your communities on #January20!

Y a aussi la mise en situation professionnelle…suis déjà au bout de ma vie avant commencer cette campagne PU😭😂

The Mathematics of Artificial Intelligence: In this introductory and highly subjective survey, aimed at a general mathematical audience, I showcase some key theoretical concepts underlying recent advancements in machine learning. arxiv.org/abs/2501.10465

It’s preprint day! 😌
Large deviations for sticky-reflecting Brownian motion with boundary diffusion - Archive ouverte HAL hal.science/hal-04895784

Since on Monday I will have a paper in probability on Arxiv, can I officially apply also for full professorship in probability?🤣🤣🤣🤣

Soft(arg)max function transforms logits into probability vector. One shows that softmax (w/ temperature) converges pointwise to the argmax, selecting the most likely outcome. Essential in in multinomial logistic regression and for attention mechanisms

Slides for a general introduction to the use of Optimal Transport methods in learning, with an emphasis on diffusion models, flow matching, training 2 layers neural networks and deep transformers. speakerdeck.com/gpeyre/optim...

J’ai terminé mon service d’enseignement mais pas de pause car il faut rédiger les dossiers prof…en fait suis un hamster enseignements->dossiers->enseignements->… 🐹

Join the PIMS Kantorovich Initiative at the CRM-PIMS-AARMS Special Session on Optimal Transport on Jan 10 & 11.

This session features talks by scientific leaders on optimal transport and its applications.

📍 Seattle Convention Center
🔗 Learn more and register: tinyurl.com/tkshh2nj

#JMM2025